Home Artists Posts Import Register

Content

Hello all Patrons!

I know it has been difficult to search through over 120+ videos I have created at echohive channel. That is why I have built https://www.echohive.live/ where you can instantly search all videos with links to code downloads.

If you feel like sharing it with the world(on social media) please send me a message about here or at discord or comment under this post and I will share the source code for the GPT-4 chatgpt UI replica with you. You can test it here https://gpt-4.up.railway.app/ !!be mindful of token usage as it is set up to only use GPT-4!!

I hope this will benefit everyone in their learning journey.

Also don't forget the upcoming AI workshop and round table discussions this Sunday. You can find the schedule here:
https://www.patreon.com/posts/3-upcoming-ai-q-81663644

PS: the thumbnail image showing for the website is incorrrect. I will work on to fix it

Files

Video Gallery

In this video we will explain how to use the LAMA Index for building composable graphs. These graphs allow you to build indexes on top of other indexes, which is beneficial when you want to query multiple indexes at the same time, something that would be difficult with traditional single index similarity searches due to the limit of chunks that can be fed into the large language model.

Comments

david benjamin

any chance you'll show us how you did this? I have a 2200 vid transcripts i need to be able to provide some kind of sesarch mechanism for. my thought was to try to download them all and embed them in a vectordb... is that what you've done?

echohive42

I am thinking of making a video on how I built it. The search is actually just regular keyword search currently. But I am thinking of implementing vector search in the future. Keep in mind that regular search is blazing fast and most of the time sufficient.

david benjamin

using what, youtube_data api or programmable search engine, or...? so its just searching meta and titles, right? not content ?

echohive42

It gets the video information with official youtube API v3 and transcriptions with youtube-transcript-api